AI and prescriptive analytics enable CPG companies to adopt an agile approach to New Product Development, thereby allowing them to react quicker to the different changes in the market. They also allow for seamless data flow by merging R&D and Marketing activities (read more about it in our New Product Development Playbook blog post.)
For CPG companies specifically, this has been particularly disruptive as the shifts in both the consumer landscape and the competitive one have forced companies to adapt or die.
Machine learning has changed how we think about marketing, product development, customer service, data analytics, and more – creating a fair amount of growing pains along the way.
The challenges here are threefold. Firstly, most CPG companies simply do not have access to the sort of granular data across their channels and markets that could feed machine learning algorithms appropriately. Secondly, their data management is suboptimal – owing in no small part to decentralized and matrixed structures. And lastly, their fragmented systems landscape is a Frankenstein monster of legacy systems that do not integrate well together.
While some companies have tried to navigate these challenges and build this AI capability in-house (we will talk about this in a future post), the far more achievable solution (at least in the short term) has been to outsource the AI-related tasks to a vendor who already has the skills, resources, and experience to leverage.
However, this sort of outsourcing comes with its own pros and cons and, in this article, we will explore some of the considerations that should be front of mind when you consider outsourcing your AI work to a third party.
Key Considerations When Outsourcing to a 3rd party vendor
When you consider outsourcing your AI function to a 3rd party vendor, you need to consider both the pros and cons so that you can make an informed decision. In that vein, let us look at some of the key benefits and risks that come with this sort of outsourcing:
|Access to Skills. Working with an AI vendor allows you to leverage key technical skills that you might not have in-house, boosting the chances of your project’s success. A third party who comes in without any preconceptions can offer a fresh perspective on your company and the challenges it faces. |
This can be invaluable for problem solving and for injecting new life into the organization
|Quicker to Launch. Building in-house capacity can take time because having to learn/build a completely new skillset in-house and provide tools that cater to your business needs can be a monumental challenge. Building the right toolset can be a huge investment, both in terms of time and money. |
Additionally, hiring is challenging and it is not always possible to find the right people. Working with an external vendor who already has the resources allows you to start quicker which makes the process more agile.
|Scalable Cost. When you outsource, you can match your cost profile to the exact projects that you are working on, rather than having to take on full-time salaries when you are not sure how effectively they will be utilized.|
|Misaligned Expectations. As with any outsourced function, it relies on clear communication and good understanding between the parties if it is going to work. If the expectations are not in sync, you risk a lot of wasted time, effort, and resources fixing things later.|
|Integration and Implementation Risk. AI projects are notorious for working in a test environment but then falling apart when pushed into production and integrated with the rest of the organization. |
Outsourcing your AI function faces this risk because the external party is not familiar with your company’s unique workflows, staff, integration needs, and the like.
|Legal and Reputational Risk. Inviting a third party to interact with your internal data brings its own legal and reputational risk especially if that information is sensitive. By relinquishing control here, you are relying on their own security measures and the confidentiality of their staff.|
You should be aware of all the above when deciding on whether to outsource or not. It is also a good idea to look into going with a vendor that has a good track record because often they can manage the risks more robustly while still giving you the leverage of the skills you seek.
Should You Use One Vendor or Multiple Vendors?
AI projects in CPGs entail many different moving parts and when you consider how they should interact with the other systems in your company, it can add up quickly. Because the specialization is seen as a competitive advantage, many companies look to engage with multiple providers who have expertise in each specific component.
This provides benefits in terms of diversification and creates a natural competition between suppliers to provide a better service. But what often gets overlooked is the hidden costs that come with having multiple vendors.
When you are working with too many different suppliers, you lose control over mission-critical functions that you should be holding onto. In addition, the costs and headaches that come with having to integrate all of those different workflows, technology stacks, and ways of doing things into one coherent process can be immense.
To put this into perspective, only 35% of IT project implementations are successful, even though most of the issues could be prevented at minimal cost by asking for the support of the software provider.
Partnering with new technology vendors requires a significant amount of time and resources to administer the relationship, implement the solution, and work through bugs and problems with each one. In all of these situations, the bottlenecks that come with having too many providers can be crippling – especially when you consider the burdens of training and support. Additionally, integrating the data that is used and produced by multiple providers can become a big challenge for organizations.
Instead, it is often better to work with one general-purpose vendor who can offer all the functionality that you need under one roof. Data from Forrester shows that 31% of survey respondents say that their efforts to gain access to specialized skills through multi-sourcing were highly successful, and 27% saying that they were successful in reducing costs thanks to encouraging competition between suppliers.
What you lose in niche specialization, you gain in the practicality of an end-to-end solution that allows you to manage the risks we mentioned in the previous section that much more precisely. You will have a more streamlined integration process, a deeper support relationship, and the ability to homogenize your data collection which aids with future machine learning projects.
How To Choose the Right Vendor?
Choosing the right technology partner is going to come down to a few things: the scope of the project, your available budget, the capacity of your internal team, and the vision for what you are trying to create. Take your time with each potential supplier and evaluate their capabilities from a practical perspective. You must be confident that they not only have the skills and expertise to assist, but also that they have a track record that points to their success in the space.
There are lots of good resources that can help with this process (See here and here) but at the crux of it, you are looking for a multi-purpose provider that can service all of your needs through one solution, essentially someone who:
- Provides technology or a platform whose functionality aligns with your business objectives and delivers business value. This will ensure senior management buy-in.
- Delivers on your business requirements. You would first need to determine whether your business can use an off-the shelf or whether it requires a customized solution.
- Possesses sufficient knowledge of your industry and the capabilities to deliver on the challenges you face. You would want to look for AI expertise within the vendor organization who would be able to assist you in solving some of your key business problems.
- Has a proven track record. Different social proof points such as success stories, testimonials and case studies can be used to determine this factor.
Ideally, you want to look at other similar projects that they have completed to understand what they are capable of and how they might help you. Lastly, you want to look for a firm that is going to work alongside you on the journey and see themselves as a partner, rather than just a supplier.
This subtle difference in attitude makes a big difference over the long term – especially with AI projects that are intricate and far-reaching.
Do not rush the process – especially when you are looking for a multi-purpose solution. The more due diligence you do upfront, the fewer problems you will encounter along the way.
Choosing the right technology vendor for your AI outsourcing is always going to depend on your context, but it is worth considering a multi-purpose software solution from one provider to simplify your implementation and concentrate your resources in one place. Companies can save a lot of time and money doing this, especially in a complicated field such as AI. The practical benefits speak for themselves and often it is this one decision that can completely transform your digital solutions.
We would be remiss if we did not mention that here at Foodpairing, we offer a sophisticated end-to-end solution for new product development utilizing the latest in artificial intelligence. If you are a large CPG company that wants to explore this, be sure to get in touch today, and let us see how we can help.